Compare|Favorites
▨ thegpucatalog.com

CMP 170HX 8 GB

General
ManufacturerNVIDIA
SeriesMining GPUs
GPU NameGA100
ArchitectureAmpere
GenerationMining GPUs
Release DateSep 2021
Bus InterfacePCIe 1.0 x4
Process & Die
FoundryTSMC
Process Size7 nm
Transistors54,200 M
Transistor Density65,600 K/mm²
Die Size826 mm²
Chip PackageBGA-2743
Clock Speeds
Base Clock1,140 MHz
Boost Clock1,410 MHz
Memory
VRAM8 GB
Memory TypeHBM2e
Memory Clock1,458 MHz
Memory Bus4096-bit
Bandwidth1,490 GB/s
Processing Units
Shading Units4,480
TMUs280
ROPs128
SMs70
Tensor Cores280
RT Cores0
Cache
L1 Cache192 KB
L2 Cache8 MB
Power & Thermals
TDP250 W
Suggested PSU600 W
Power Connectors2x 8-pin
Dimensions
Outputs
API Support
OpenCL3.0
CUDA8.0
Theoretical Performance
Pixel Rate180.5 GP/s
Texture Rate394.8 GT/s
FP16 (Half)50,530 GFLOPS
FP32 (Single)12,630 GFLOPS
FP64 (Double)6,317 GFLOPS
External
TechPowerUpView →